Add some convenience/helper functions for RSA.

Add some convenience/helper functions for RSA. Modify test utility to use the new function.

BUG=670
TEST=RSA verification test using the convenience function is passes.

[sosa, 2010-02-05 19:16:45]

In general, you may want to change all your include files to include from
vboot_reference for clarity.  If you look through pam_google and window_manager
under platform, they include all from the head directory.  If you run lint, lint
also asks that you do this.  Not super necessary, but would be nice.  Could
definitely wait for another cl to do it.

http://codereview.chromium.org/575019/diff/1001/1003
File src/platform/vboot_reference/crypto/rsa_utility.c (right):

http://codereview.chromium.org/575019/diff/1001/1003#newcode8
src/platform/vboot_reference/crypto/rsa_utility.c:8: #include "padding.h"
You should use paths relative to your base directory in your includes to avoid
confusion of where these headers are.  For example
vboot_reference/include/padding.h versus padding .h and have your Makefile
include from vboot_reference

http://codereview.chromium.org/575019/diff/1001/1003#newcode13
src/platform/vboot_reference/crypto/rsa_utility.c:13: int key_len =
siglen_map[algorithm] * sizeof(uint32_t);  /* Key length in
You should try to use // comment-style for multi-line comments like this

http://codereview.chromium.org/575019/diff/1001/1003#newcode15
src/platform/vboot_reference/crypto/rsa_utility.c:15: return (2 * key_len +
sizeof(int) + sizeof(uint32_t));
Can you explain this math in a comment.  What is the int and the uint32?  Why
*2?

[gauravsh, 2010-02-05 19:34:26]

http://codereview.chromium.org/575019/diff/1001/1003
File src/platform/vboot_reference/crypto/rsa_utility.c (right):

http://codereview.chromium.org/575019/diff/1001/1003#newcode8
src/platform/vboot_reference/crypto/rsa_utility.c:8: #include "padding.h"
On 2010/02/05 19:16:45, sosa wrote:
> You should use paths relative to your base directory in your includes to avoid
> confusion of where these headers are.  For example
> vboot_reference/include/padding.h versus padding .h and have your Makefile
> include from vboot_reference

This is supposed to be a standalone self-contained reference implementation. It
won't be linked into a Chrome OS build. Most likely, I will move it to its own
git repo at some point. Do you think base relative paths are still necessary in
that case?

http://codereview.chromium.org/575019/diff/1001/1003#newcode13
src/platform/vboot_reference/crypto/rsa_utility.c:13: int key_len =
siglen_map[algorithm] * sizeof(uint32_t);  /* Key length in
On 2010/02/05 19:16:45, sosa wrote:
> You should try to use // comment-style for multi-line comments like this

As this is for use in the firmware, it's entirely written in ANSI C. Hence, not
using '//'.

http://codereview.chromium.org/575019/diff/1001/1003#newcode15
src/platform/vboot_reference/crypto/rsa_utility.c:15: return (2 * key_len +
sizeof(int) + sizeof(uint32_t));
On 2010/02/05 19:16:45, sosa wrote:
> Can you explain this math in a comment.  What is the int and the uint32?  Why
> *2?

Ok, this is related to the RSAPublicKey structures (which holds the
pre-processed RSA public key). Added a comment.

[sosa, 2010-02-05 19:37:59]

lgtm.  No if it's stand-alone that's fine

http://codereview.chromium.org/575019/diff/5001/3005
File src/platform/vboot_reference/tests/verify_data.c (right):

http://codereview.chromium.org/575019/diff/5001/3005#newcode82
src/platform/vboot_reference/tests/verify_data.c:82: uint8_t* digest = NULL;
Is this extra whitespace?

[gauravsh, 2010-02-05 19:40:45]

Pushing...

http://codereview.chromium.org/575019/diff/5001/3005
File src/platform/vboot_reference/tests/verify_data.c (right):

http://codereview.chromium.org/575019/diff/5001/3005#newcode82
src/platform/vboot_reference/tests/verify_data.c:82: uint8_t* digest = NULL;
On 2010/02/05 19:37:59, sosa wrote:
> Is this extra whitespace?
No, I think it's just the diff, as I split the declaration to separate lines.